急性呼吸道感染的病毒检测新策略:基于SERS和机器学习方法  

A Novel Strategy for Viral Detection in Acute Respiratory Infections:Combining SERS With Machine Learning

作　　者：姜恒 吕紫薇 李洋 董妥 JIANG Heng;LÜZi-wei;LI Yang;DONG Tuo(School of Public Health,Harbin Medical University,Harbin 150081,China;School of Pharmacy,Harbin Medical University,Harbin 150081,China)

机构地区：[1]哈尔滨医科大学公共卫生学院,黑龙江哈尔滨150081 [2]哈尔滨医科大学药学院,黑龙江哈尔滨150081

出　　处：《光谱学与光谱分析》2025年第5期1217-1224,共8页Spectroscopy and Spectral Analysis

基　　金：国家自然科学基金项目(32000137)资助。

摘　　要：快速准确地检测引发急性呼吸道感染(ARI)的常见病毒,对于公共卫生防控至关重要。尽管传统的病毒检测方法在一定程度上满足了临床需求,但其往往存在耗时较长、成本较高或灵敏度有限等局限性,亟需更为快速和高效的检测手段。表面增强拉曼光谱(SERS)技术因其高灵敏度和特异性,逐渐成为病毒检测领域的研究热点。研究旨在开发一种结合SERS技术与机器学习方法的新型高效检测策略,以实现对呼吸道合胞病毒(RSV)、甲型流感病毒(IFA)和人腺病毒(HAdV)的精准检测。采用柠檬酸盐制备银纳米颗粒(Ag@cit),将碘离子孵育和钙离子聚集的银纳米颗粒(Ag@ICNPs)作为SERS基底。Ag@ICNPs具有适合病毒检测的优质“热点”,能够超快速、高灵敏、无标记地捕获呼吸道病毒的特征指纹图谱。为了进一步提高检测的效率和准确性,将机器学习方法引入到SERS技术中,通过对多种机器学习算法的改进,成功建立了病毒分类器,能够在3 min内对检测限低至1.0×10^(2)copies·mL^(-1)的三种病毒进行快速鉴定,且准确率高达100%。此外,利用病毒浓度与特征峰强度间关系所构建的浓度依赖性曲线,具有良好的线性关系(R2均大于0.998),为量化样本中病毒含量提供了可能性,这对于临床通过病毒负荷的变化监测治疗效果和病情进展具有重要意义。该研究揭示了“SERS@机器学习”联合应用在呼吸道病毒快速精准检测中的显著优势,并为ARI临床诊断提供了一种具有潜在应用价值的新途径,有望在未来成为临床诊断和公共卫生防控中的重要工具。Rapid and accurate detection of common viruses causing acute respiratory infections(ARI)is crucial for public health prevention and control.Although traditional viral detection methods have partially met clinical needs,they often have limitations such as long detection times,high costs,or limited sensitivity.There is an urgent need for faster and more efficient detection methods.Surface-Enhanced Raman Spectroscopy(SERS)has become a research hotspot in viral detection due to its high sensitivity and specificity.This study aims to develop a novel and efficient detection strategy combining SERS technology with machine learning methods to achieve precise detection of Respiratory Syncytial Virus(RSV),Influenza A Virus(IFA),and Human Adenovirus(HAdV).The study employs citrate-stabilized silver nanoparticles(Ag@cit)and uses iodine ion incubation and calcium ion aggregation to prepare silver nanoparticles(Ag@ICNPs)as the SERS substrate.Ag@ICNPs have high-quality“hotspots”suitable for virus detection,enabling ultra-fast,highly sensitive,and label-free capture of characteristic fingerprint spectra of respiratory viruses.This study integrates machine learning methods with SERS technology to further improve detection efficiency and accuracy.By improving various machine learning algorithms,a virus classifier was successfully established,which can rapidly identify the three viruses within 3 minutes with a detection limit as low as 1.0×10^(2)copies·mL^(-1)and an accuracy rate of 100%.Additionally,the concentration-dependent curves constructed based on the relationship between viral concentration and characteristic peak intensity showed good linearity(R2 greater than 0.998),providing the possibility for quantifying virus content in samples.This is important for monitoring treatment efficacy and disease progression through changes in viral load in clinical settings.This study reveals the significant advantages of the combined application of“SERS@machine learning”in rapidly and precisely detecting respiratory viruses,off

关 键 词：表面增强拉曼光谱 机器学习 呼吸道病毒 无标记检测 

分 类 号：O433[机械工程—光学工程]